An online detection method for inter-turn short circuit faults in permanent magnet motors

Abstract

The present invention discloses an online detection method for inter-turn short circuit faults of permanent magnet motors, which relates to the field of fault detection of permanent magnet motors. The present invention mainly focuses on the online fault diagnosis of permanent magnet synchronous motors when inter-turn short circuits occur. First, a mathematical model of inter-turn short circuit faults and the characteristic quantities generated after the faults occur are derived. The most direct characteristic quantities are the short-circuit current and the voltage disturbance quantity during the fault. The actual inter-turn short circuit loop current and the fault voltage disturbance quantity cannot be directly measured, and the voltage disturbance quantity is extracted by constructing a state observer. The entire system consists of three parts, namely, a voltage disturbance quantity observation module based on LESO, a fault characteristic extraction algorithm module, and a fault diagnosis method module. The present invention overcomes the calculation defects of multiple patent applications in the background art, can perform online calculation on the existing TCU control device, can provide real-time protection when the motor has an inter-turn short circuit, improves the system safety; and ensures the real-time performance and reliability of the inter-turn short circuit fault diagnosis of the motor.

Description

Technical Field

[0001] The present invention relates to the field of fault detection of permanent magnet motors, and specifically to an online detection method for inter-turn short circuit faults of permanent magnet motors. Background Technique

[0002] As an important core component of the rail transit system, the reliability of the permanent magnet motor will affect the performance of the entire system. The operating conditions of the rail transit traction motor are complex and changeable. Due to the influence of rail surface vibration shock, frequent instantaneous overload and shock, humid and high-temperature environment, etc., it is easy for the traction motor to fail. Among the entire motor faults, the inter-turn short circuit fault accounts for about 36% of the motor faults. However, before a serious inter-turn short circuit fault occurs, a series of deterioration signs such as vibration noise and electromagnetic harmonics will appear. If the inter-turn short circuit fault can be detected early through fault diagnosis technology, the safety of rail transit can be greatly improved.

[0003] At present, the inter-turn short-circuit fault detection methods mainly rely on big data signal analysis, deep learning, and intelligent algorithms. These signal analysis and intelligent algorithms have a large computational workload and poor real-time performance, making it difficult to achieve online fault diagnosis. For example: ① Patent CN101221206A performs instantaneous power analysis on the motor winding during operation, uses the first-order complex Gaussian wavelet function to extract fault features from the instantaneous power distortion points before and after the fault, and determines whether the motor has an inter-turn short circuit; ② Patent CN111208427A calls a preset gradient boosting tree model and a preset fuzzy clustering model, respectively uses spectrum analysis, Park vector analysis, and symmetrical component analysis methods to extract their respective fault features, forms a new multi-dimensional feature vector, and extracts feature quantities after dimensionality reduction for inter-turn short-circuit fault detection; ③ Patent CN108107315A collects the current and control signals of each phase of the three-phase winding of a salient-pole permanent magnet motor in normal and short-circuit conditions to establish normal and fault models, and compares the evaluation result of the winding severity with the obtained immunity fault severity parameter and the preset parameter range; ④ Patent CN110988674A is a method and system for monitoring the health status of a permanent magnet synchronous motor based on multi-modal canonical correlation analysis. Select the normal operation data of the permanent magnet synchronous motor under two or more working conditions as the training data set, input the EM algorithm to calculate the parameters required for modeling, obtain the online operation data of the permanent magnet synchronous motor, and detect the health status of the permanent magnet motor by combining the evaluation index obtained from the model and the preset fault probability threshold; ⑤ Patent CN111123105A is a method for diagnosing inter-turn short-circuit faults in motors based on high-frequency signal injection. Inject sinusoidal high-frequency currents with equal amplitudes and a phase difference of 180° into the winding with an inter-turn short-circuit fault, extract the frequency components identical to the injected high-frequency from the two-phase voltage, calculate the fault feature quantity, and determine whether an inter-turn short-circuit fault has occurred based on the comparison result between the feature quantity and the set threshold. The first four patent applications are based on methods of wavelet Gaussian, FFT signal analysis, and intelligent learning, which require a large amount of offline data storage and analysis to extract fault features and are not easy to achieve online real-time data calculation for fault monitoring; the fifth patent application is also an offline monitoring method for diagnosing inter-turn short-circuit faults in motors by high-frequency signal injection. The motor is monitored for faults when it is stationary and stopped. The switching frequency of rail transit is generally only a few hundred hertz, which limits the injection frequency of high-frequency signals.

[0004] Therefore, based on the above problems, it is necessary to design a new fault monitoring method or improve the existing fault monitoring method. Summary of the Invention

[0005] In order to solve the problems of the existing inter-turn short-circuit fault detection methods, such as large computational workload of signal analysis and intelligent algorithms, poor real-time performance, and difficulty in achieving online fault diagnosis, the present invention provides a new online fault detection method for inter-turn short circuits of permanent magnet motors based on state observation.

[0006] The present invention mainly focuses on the on-line fault diagnosis of turn-to-turn short circuits in permanent magnet synchronous motors. First, a mathematical model of turn-to-turn short circuit faults and the characteristic quantities generated after the faults occur are derived. The most direct characteristic quantities are the short-circuit current and the voltage disturbance during the fault. The actual turn-to-turn short circuit loop current and the fault voltage disturbance cannot be directly measured. By constructing a linear extended state observer, the voltage disturbance of the motor after turn-to-turn short circuit during operation is observed, and through signal filtering processing, threshold comparison is performed to determine whether the motor has a turn-to-turn short circuit fault. Figure 1 A system block diagram of the on-line fault diagnosis method for turn-to-turn short circuits in PMSM is given. The entire system consists of three parts, namely, a voltage disturbance observation module based on LESO, a fault feature extraction algorithm module, and a fault diagnosis method module.

[0007] The present invention is realized through the following technical solutions: An on-line detection method for turn-to-turn short circuit faults in permanent magnet motors, specifically including the following steps:

[0008] 1) Assume that the A-phase winding in the three-phase winding has a turn-to-turn short circuit, and establish the d-axis and q-axis voltage equations after the turn-to-turn short circuit as follows: When the B-phase and C-phase have faults, all θ in the formula are respectively changed to (θ - 120°) or (θ + 120°);

[0009]

[0010] In the formula:

[0011]

[0012] Among them, L s0 , M s0 DC components on self-inductance and mutual inductance; L s2 Amplitude of the sine component on the inductance; η is the short-circuit turn ratio, R f Is the short-circuit resistance, L d Is the d-axis inductance, L q Is the q-axis inductance, R s Is the phase resistance of the motor; ψ f Is the magnitude of the permanent magnet flux linkage; θ is the angle between the d-axis and the a-axis; p is the differential operator;

[0013] Given the voltage commands u d And u q , the rotational speed is ω, then:

[0014]

[0015] u an = u a - u n , u an Is the voltage between phase A and the actual center point N of the star-connected motor;

[0016] Substitute formula (2) into to obtain:

[0017]

[0018] 2) Observe that the definition related to the fault in formula (1) is the voltage disturbance term, and the voltage disturbance term is one of the most direct characteristics of the inter-turn short-circuit fault:

[0019]

[0020] Substitute formula (3) into formula (4) and simplify to obtain:

[0021]

[0022] In the formula, e df , e qf are the voltage disturbance quantities after the d-axis and q-axis turn shorts; η is the shorted turn ratio, R f is the short-circuit resistance; L d , L q , R s are the motor parameters; β, δ1 = arctan(-R s / ω(2L d -L q ), δ2 = arctan(-ω(L d -2L q ) / R s ) are all constants; from formula (5), it can be seen that the voltage disturbance quantity is divided into a DC component and a second harmonic component;

[0023] 3) In steps 1) and 2), the modeling process of the inter-turn short-circuit fault and the characteristic quantities generated after the fault are derived. The most direct characteristic quantities are the short-circuit current and the fault voltage disturbance quantity. In actual situations, the short-circuit current cannot be directly measured, while the voltage disturbance quantity can be observed according to the motor voltage equation. The motor voltage equation is as follows:

[0024]

[0025] Among them, e df is the voltage disturbance quantity of the d-axis after the turn short; e qf is the voltage disturbance quantity of the q-axis after the turn short;

[0026] 4) Construct the d-axis and q-axis disturbance voltage LESO disturbance observers according to the permanent magnet motor voltage transient equation and the principle of the linear extended state observer LESO as follows:

[0027]

[0028]

[0029] When there is no turn - to - turn short - circuit fault in the motor, When there is a turn - to - turn short - circuit fault in the motor,

[0030] 5) It is found through analysis that the voltage disturbance obtained by the observer is equivalent to the actual disturbance;

[0031] 6) Selection and observation of voltage disturbance characteristic quantities. The voltage disturbance quantity is divided into a DC component and a second - harmonic component. The formula is as follows:

[0032]

[0033] 7) Select the second - harmonic component of the voltage disturbance quantity as the characteristic quantity for turn - to - turn short - circuit. Perform positive - and negative - sequence Park transforms on the second component. After the transformation, there are DC components, second components, and fourth components. Design a low - pass filter to extract the DC component of the system as the basis for threshold judgment:

[0034] Perform positive - sequence Park transform on the second component:

[0035] Perform low - pass filtering on the above formula:

[0036] Perform negative - sequence Park transform on the second component:

[0037] Perform low - pass filtering on the above formula:

[0038] 8) Final characteristic extraction observation threshold:

[0039]

[0040] Ignoring the resistance, when the salient - pole ratio L q / L d > 2, A > B; when the salient - pole ratio L q / L d < 2, A < B; when L q / L d = 1, A = 0; Therefore, the observed characteristic quantity ensures applicability to motors with different salient - pole ratios;

[0041] 9) Finally, use the threshold Th as the criterion for judging whether a fault occurs. The threshold Th is determined according to the values of the digital and semi - physical simulation observations Index of the system under full - load and no - load conditions, and the threshold Th is corrected considering non - linear factors during actual tests; when the Index calculated by formula (10) exceeds the threshold Th, it is considered that a turn - to - turn short - circuit fault has occurred in the motor, and the corresponding system performs a protection action; conversely, when Index is less than the threshold Th, it is considered that the motor is operating normally.

[0042] Compared with the prior art, the present invention has the following beneficial effects: The on-line detection method for inter-turn short circuit fault of a permanent magnet motor provided by the present invention can perform real-time diagnosis on whether an inter-turn short circuit fault occurs in a permanent magnet synchronous motor, overcoming the drawbacks of the prior art methods in patents CN110988674A and CN111123105A that cannot perform real-time on-line diagnosis; it does not require additional detection equipment, and uses a linear extended state observer (LESO) to estimate the voltage disturbance amount after the turn short, and determines whether an inter-turn short circuit fault occurs, ensuring the real-time performance and reliability of the diagnosis of the inter-turn short circuit fault of the motor. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a system block diagram of the on-line fault diagnosis method for inter-turn short circuit of PMSM. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0044] The present invention will be further described below in conjunction with specific embodiments.

[0045] Figure 1 A system block diagram of the on-line fault diagnosis method for inter-turn short circuit of PMSM is given. The whole system consists of three parts, namely a voltage disturbance amount observation module based on LESO, a fault feature extraction algorithm module, and a fault diagnosis method module.

[0046] An on-line detection method for inter-turn short circuit fault of a permanent magnet motor specifically includes the following steps:

[0047] 1) Assume that the A-phase winding in the three-phase winding has an inter-turn short circuit, and establish the voltage equations of the d-axis and q-axis after the turn short as follows: When the B-phase and C-phase have faults, all θ in the formula are respectively changed to (θ - 120°) or (θ + 120°);

[0048]

[0049] In the formula:

[0050]

[0051] Among them, L s0 , M s0 DC components on self-inductance and mutual inductance; L s2 Amplitude of the sine component on the inductance; η is the short-circuit turn ratio, R f is the short-circuit resistance, L d is the d-axis inductance, L q is the q-axis inductance, R s is the phase resistance of the motor; ψ f is the magnitude of the permanent magnet flux linkage; θ is the angle between the d-axis and the a-axis; p is the differential operator;

[0052] Given the voltage command u d and u q, with a rotational speed of ω, then:

[0053]

[0054] u an = u a - u n where u an is the voltage between phase A and the actual center point N of the star-connected motor;

[0055] Substitute formula (2) into to get:

[0056]

[0057] 2) Observe that the definition related to the fault in formula (1) is the voltage disturbance term, and the voltage disturbance term is one of the most direct characteristics of the turn-to-turn short circuit fault:

[0058]

[0059] Substitute formula (3) into formula (4) and simplify to get:

[0060]

[0061] In the formula, e df , e qf are the voltage disturbance amounts after turn-to-turn short circuit on the d and q axes; η is the shorted turn ratio, R f is the short circuit resistance; L d , L q , R s are the motor parameters; β, δ1 = arctan(-R s / ω(2L d - L q ), δ2 = arctan(-ω(L d - 2L q ) / R s ) are all constants; from formula (5), it can be seen that the voltage disturbance amount is divided into a DC component and a second harmonic component;

[0062] 3) In steps 1) and 2), the modeling process of the turn-to-turn short circuit fault and the characteristic quantities generated after the fault are derived. The most direct characteristic quantities are the short circuit current and the fault voltage disturbance amount. In actual situations, the short circuit current cannot be directly measured, while the voltage disturbance amount can be observed according to the motor voltage equation. The motor voltage equation is as follows:

[0063]

[0064] where e df is the voltage disturbance amount on the d axis after turn-to-turn short circuit; e qfIt is the disturbance quantity of the q-axis voltage after turn-to-turn short circuit;

[0065] 4) Construct the LESO disturbance observers for the d-axis and q-axis disturbance voltages according to the voltage transient equation of the permanent magnet motor and the principle of the linear extended state observer LESO as follows:

[0066]

[0067]

[0068] When no turn-to-turn short circuit fault occurs in the motor, When a turn-to-turn short circuit fault occurs in the motor,

[0069] 5) Through analysis, it is found that the voltage disturbance obtained by the observer is equivalent to the actual disturbance;

[0070] 6) Selection and observation of voltage disturbance characteristic quantities. The voltage disturbance quantity is divided into a DC component and a second harmonic component, and the formula is as follows:

[0071]

[0072] 7) Select the second harmonic component of the voltage disturbance quantity as the characteristic quantity during turn-to-turn short circuit. Perform positive and negative sequence Park transforms on the second component. After the transformation, there are DC components, second components, and fourth components. Design a low-pass filter to extract the DC component of the system as the basis for threshold judgment:

[0073] Perform positive sequence Park transform on the second component:

[0074] Perform low-pass filtering on the above formula:

[0075] Perform negative sequence Park transform on the second component:

[0076] Perform low-pass filtering on the above formula:

[0077] 8) Final characteristic extraction observation threshold:

[0078]

[0079] Ignoring the resistance, when the salient pole ratio L q / L d > 2, A > B; when the salient pole ratio L q / L d < 2, A < B; when L q / L d = 1, A = 0; Therefore, the observed characteristic quantity ensures applicability to motors with different salient pole ratios;

[0080] 9) Finally, the threshold Th is used as the criterion for judging whether a fault occurs. The threshold Th can be determined according to the values of the digital and semi-physical simulation observations Index under full load and no load of the system, and the threshold Th is corrected considering non-linear factors during actual tests; when the Index calculated by formula (10) exceeds the threshold Th, it is considered that a turn-to-turn short circuit fault has occurred in the motor at this time, and the corresponding system performs a protection action; on the contrary, when Index is less than the threshold Th, it is considered that the motor is operating normally.

[0081] The scope of protection required by the present invention is not limited to the above specific embodiments, and for those skilled in the art, the present invention can have various deformations and changes. Any modifications, improvements, and equivalent replacements made within the concept and principle of the present invention should be included within the scope of protection of the present invention.

Claims

1. An on-line detection method for inter-turn short circuit faults of a permanent magnet motor, characterized in that: This method is implemented in three modules, namely the LESO voltage disturbance observation module, the fault feature extraction algorithm module, and the fault diagnosis method module. The specific steps are as follows: 1) Assume that a turn-to-turn short circuit occurs in the A-phase winding of the three-phase winding. The d-axis and q-axis voltage equations after the turn-to-turn short circuit are established as follows: When a fault occurs in the B-phase or C-phase, all θ in the formula are changed to (θ - 120°) or (θ + 120°) respectively; In the formula: Among them, L s0 , M s0 DC components on self-inductance and mutual inductance; L s2 Amplitude of the sinusoidal component on the inductor; η is the short-circuit turn ratio, R f is the short-circuit resistance, L d is the d-axis inductance, L q is the q-axis inductance, R s is the motor phase resistance; ψ f is the magnitude of the permanent magnet flux linkage; θ is the angle between the d-axis and the a-axis; p is the differential operator; Given voltage command u d and u q , with the rotational speed being ω, then: u an = u a - u n ,u an is the voltage between phase A and the actual center point N of the star-connected motor; Substitute formula (2) into to obtain: 2) Observe that the term related to the fault in formula (1) is defined as the voltage disturbance term, which is one of the most direct characteristics of the turn-to-turn short circuit fault: Substitute formula (3) into formula (4) and simplify to get: where, e df , e qf are the voltage disturbance quantities after the short - circuit of the turns on the d - and q - axes; η is the short - circuit turn ratio, R f is the short - circuit resistance; L d , L q , R s are motor parameters; δ1 = arctan(-R s / ω(2L d - L q ))), δ2 = arctan(-ω(L d - 2L q ) / R s ) are both constants; It can be seen from Equation (5) that the voltage disturbance quantity is divided into a DC component and a second - harmonic component; 3) Steps 1) and 2) deduce the modeling process of the turn-to-turn short circuit fault and the characteristic quantities generated after the fault occurs. The most direct characteristic quantities are the short-circuit current and the fault voltage disturbance. In actual situations, the short-circuit current cannot be directly measured, while the voltage disturbance can be observed according to the motor voltage equation. The motor voltage equation is as follows: where, e df is the d-axis voltage disturbance amount after turn-to-turn short circuit; e qf is the q-axis voltage disturbance amount after turn-to-turn short circuit; 4) Construct the d-axis and q-axis disturbance voltage LESO disturbance observers according to the permanent magnet motor voltage transient equation and the principle of the linear extended state observer LESO, as follows: When the motor does not have a turn-to-turn short circuit fault, When the motor has a turn-to-turn short circuit fault, 5) Through analysis, it is found that the voltage disturbance obtained by the observer is equivalent to the actual disturbance; 6) Selection and observation of voltage disturbance characteristic quantities. The voltage disturbance is divided into a DC component and a second harmonic component. The formula is as follows: 7) Select the second harmonic component of the voltage disturbance as the characteristic quantity for turn-to-turn short circuit. Perform positive and negative sequence Park transforms on the second component. After the transformation, there are DC components, second components, and fourth components. Design a low-pass filter to extract the DC component of the system as the basis for threshold judgment: Perform the positive sequence Park transformation on the secondary component: Perform low-pass filtering on the above formula: Perform the negative-sequence Park transformation on the secondary component: Perform low-pass filtering on the above formula: 8) Final characteristic extraction observation threshold: Ignoring the resistance, when the salient pole ratio L q / L d > 2, A > B; when the salient pole ratio L q / L d < 2, A < B; when L q / L d = 1, A = 0; Therefore, the observed characteristic quantity ensures applicability to motors with different salient pole ratios; 9) Finally, use the threshold Th as the criterion for judging whether a fault has occurred. The threshold Th is determined according to the value of the Index observed in the full-load, no-load digital and semi-physical simulations of the system. During actual tests, the threshold Th is corrected considering non-linear factors; when the Index calculated by formula (10) exceeds the threshold Th, it is considered that a turn-to-turn short circuit fault has occurred in the motor at this time, and the corresponding system performs a protection action; conversely, when the Index is less than the threshold Th, it is considered that the motor is operating normally.

Images